Deep learning for short-term traffic flow prediction. (June 2017)

• Record Type: Journal Article
• Authors: Polson, Nicholas G.; Sokolov, Vadim O.
• Is Part Of: Transportation research. Volume 79(2017)
• Page Start: 1
• Page End: 17
• Publication Date: 2017-06
• Subjects: Traffic Flows -- Deep Learning -- Trend filtering -- Sparse linear models; Transportation -- Periodicals; Transportation -- Technological innovations -- Periodicals; 388.011
• DOI: 10.1016/j.trc.2017.02.024
• Languages: English
• ISSNs: 0968-090X

Abstract

Highlights: A deep learning architecture that captures nonlinear spatio-temporal flow effects. Traffic predictions during special events, a Chicago Bears football game and a snowstorm Our approacch outperforms linear and one-layer neural network models. Abstract: We develop a deep learning model to predict traffic flows. The main contribution is development of an architecture that combines a linear model that is fitted using ℓ 1 regularization and a sequence of tanh layers. The challenge of predicting traffic flows are the sharp nonlinearities due to transitions between free flow, breakdown, recovery and congestion. We show that deep learning architectures can capture these nonlinear spatio-temporal effects. The first layer identifies spatio-temporal relations among predictors and other layers model nonlinear relations. We illustrate our methodology on road sensor data from Interstate I-55 and predict traffic flows during two special events; a Chicago Bears football game and an extreme snowstorm event. Both cases have sharp traffic flow regime changes, occurring very suddenly, and we show how deep learning provides precise short term traffic flow predictions.